Exercise Scenario Setup
BASTION supports structured training exercises built around scenario packages. This guide covers scenario creation, data seeding, and exercise management.
Example: Pacific Strategy AY26
The included example scenario covers an Indo-Pacific contingency with six phases:
| Phase | Description |
|---|---|
| Competition | Gray-zone activities, diplomatic signaling, force posture adjustments |
| Crisis | Escalation triggers, decision points, coalition coordination |
| Conflict Day 4 | Initial combat operations, strike planning, defensive posture |
| Conflict Day 10 | Sustained operations, logistics strain, adaptive planning |
| Conflict Day 22 | Culmination assessment, transition planning |
| Negotiation | Ceasefire terms, force disposition, post-conflict stability |
Scenario files live in the /scenario/ directory.
Seed Scripts
Two scripts populate the database with scenario data:
# Seed the scenario (phases, documents, events, conditions)
bash scripts/seed-scenario.sh
# Seed the graph database (RAFT entities, relationships)
bash scripts/seed-graph-data.sh
Both scripts are idempotent and can be re-run safely.
Scenario Package Upload
Scenarios can also be uploaded through the UI:
- Navigate to the Understand tab in a problem set.
- Use the scenario upload area to submit a scenario package (ZIP containing documents, orders, intelligence products).
- BASTION processes the package, extracts text, and indexes documents.
AI Tag Inference
When documents are ingested, AI agents automatically categorize them:
- Documents are analyzed for content type (OPORD, INTSUM, logistics estimate, etc.).
- Tags are inferred based on doctrinal document types and content analysis.
- Tags feed into the graph database for cross-referencing and staff agent context.
Team Assignment
Exercises support role-based team assignment:
| Team | Role |
|---|---|
| Blue | Friendly force planning and execution |
| Red | Adversary/OPFOR role-play and response |
| White | Exercise control, adjudication, assessment |
Teams are assigned at the problem set level and control visibility and permissions.
Phase Gate Management
Exercise phases are managed through a Kanban board interface:
- Each phase is a column with associated tasks, decision points, and deliverables.
- Phase transitions require completion criteria (configurable per exercise).
- White team controls phase advancement.
- Phase changes trigger AI agent context updates so staff recommendations reflect the current phase.
Creating an Exercise
Use the Create from Scenario wizard (CreateProblemSetWizard):
- Select a scenario package or upload a new one.
- Configure exercise parameters (teams, timeline, phase gates).
- Assign participants to teams and staff roles.
- The wizard creates the problem set, seeds initial data, and activates AI agents.